CMV is a major disease-causing agent in transplant patients, other immunocompromised patients, and newborns. About 40,000 infants are born shedding CMV every year in the US. Of these, 8,000 are born with symptoms and/or severe handicaps and up to 8,000 more will later develop progressive hearing loss. About half of pregnant mothers have adequate immunity naturally. Thus it is known that effective mAbs exist in human blood. This is also shown by successful passive transfer of immunity by intravenously administered gamma globulin (IVIG), which has shown very high efficacy for protecting the fetus. This is in contrast to the limited efficacy observed for IVIG in the transplant setting, for which cellular immunity is apparently more important than humoral immunity.
A substantial portion of the natural response to CMV is directed towards the gB protein (Park, J. W., et al., J. Korean Med. Sci. (2000) 15:133-138). The Towne vaccine is an attenuated live virus vaccine passaged extensively in vitro, which induces antibodies that neutralize fibroblast infection, but not endothelial cell infection. This vaccine is known to be safe and has been studied for 20 years (Adler, S. P., et al., Pediatr. Infect. Dis. J. (1998) 17:200-206). Blood donors useful for isolating antibodies to gB as described below include seropositive individuals with previous exposure to CMV and seronegative subjects before and after vaccination with the Towne vaccine.
Antibodies to gB protein of CMV have been prepared (Nozawa N., et al., J. Clin. Virol. (2009) [Epub ahead of print], Nakajima, N., et al. (US2009/0004198 A1), Lanzavecchia, A, et al. (US2009/0004198 A1), Ohlin, M., et al. (J. Virol. (1993) 67:703-710). A neutralizing antibody to the AD-2 domain of gB, ITC88, has been reported (Lantto, Virology (2003) 305:201-209). However, prior efforts to clone human antibodies against CMV, while successful, are limited in scope and no high affinity (sub-nanomolar) antibodies have been described. High affinity is a key parameter as weak affinity antibodies to CMV actually promote transmission across the human placenta (Nozawa, supra), an aspect of the pathology not seen in rodents. Human CMV has a double stranded DNA genome of approximately 236 kb and is a prototypical member of the β-herpesvirus family. The high complexity of the genome means that there are many potential antigens of interest. Efforts to characterize neutralizing antibodies and their associated epitopes resulted in a subunit vaccine based on glycoprotein B (gB) that elicits an effective neutralizing response, but, when tested in a cohort of seronegative women has only 50% efficacy. This appears to be the highest efficacy of any CMV vaccine. Since vaccines typically induce antibodies with a range of affinities, the disappointing efficacy of the tested vaccines to date may be attributable to the requirement for high affinity antibodies, which argues in favor of supplying a high affinity mAb directly as a prophylactic strategy.
Failure to focus the immune response on the specific neutralizing epitopes has also been postulated as the cause of the poor efficacy (Marshall, B. C., et al., Viral. Immunol. (2003) 16:491-500. Another suspected technical problem in developing anti-CMV vaccines is that they have only been assessed for their ability to generate antibodies that neutralize fibroblast infection although infection of other cell types has increasingly become a focus for understanding the viral pathology. This bias reflects technical obstacles with regard to growth of the virus in vitro. Repeated virus passage on fibroblast cells is believed to have caused many lab strains to lose tropism for endothelial and epithelial cells. During the last few years, this deficit has been associated with the loss of one or more components of the gH/gL/UL131-UL128 glycoprotein complex on the virus surface.
Clearly a need exists for a more effective anti-CMV prophylaxis strategy.